An artificial intelligence (AI) algorithm can listen to the sound of a patient’s urination to determine the effectiveness and success of abnormal “flow” patterns along with corresponding health issues. This deep learning tool, known as Audioflow, has so far functioned almost like a specialized machine in clinics, providing similar results for patients in urology.
Dr. Lee Han Jie from Singapore General Hospital, the head of the research team, stated that there is currently a trend towards using machine learning technology in various fields because clinical doctors do not have much time to closely monitor each patient. Additionally, especially since the pandemic, there has been a shift towards telemedicine with less direct hospital care. Their research team aims to develop a method to monitor patients to assess their conditions between hospital visits.
This AI can even perform better than some healthcare staff. (Illustrative image).
The algorithm evaluates sounds generated by urine in a soundproof environment, but researchers hope to develop an application capable enough for patients to use for self-monitoring at home. The current measurement method is effective in assessing urinary-related conditions but requires patients to urinate into a machine during outpatient examinations.
The Covid-19 pandemic has limited patients’ access to clinics. The team of experts wants to develop a more effective way to assess urine at home without any medical assistance, and thus they enlisted the help of the technical department to develop a urine evaluation algorithm. To train and validate this algorithm, they recruited 534 men to participate from December 2017 to July 2019. The process was quite simple: participants used a conventional urine flow meter in a soundproof room and recorded their urination process with a smartphone.
Using only about 220 recordings, the AI learned to accurately assess flow rate, volume, and duration, all of which can indicate blockages or bladder issues. This AI can even perform better than some non-specialist healthcare staff and approach the level of senior consultants.
But the real benefit is having a consulting expert in the bathroom with you every time you urinate. The research team is now working towards making the algorithm functional even in noisy environments typical of home settings, which would be a game changer for patients.
The Audioflow device has produced results that can compete with a conventional urine meter and a panel of six residents in the urology department. The AI technology has provided conclusions consistent with standard urine flow measurements for over 80% of the recordings, and compared to urology specialists, it achieved an agreement rate of 84%. Researchers now also hope that this new AI can soon demonstrate benefits in home settings.
Audioflow will soon be launched as a smartphone application for real-world testing amidst a lot of surrounding noise. Currently, it still has a limitation as it has only been tested on male urine sounds, different from female flow. A version focused on women may be on the horizon.
